Are non-English languages supported?
Yes, text in non-English or mixed languages is supported, with certain limitations. Upload, exploration, sandbox preview, and categorization are supported for non-English languages. Sentiment analysis is currently only supported for English texts. For categorization, it is recommended to either use separate category models for separate languages, or to have one “rule” per node per language, with a AND condition restricting the rule to apply to only content of a certain language. In this mode of operation, it is recommended, though not required, to include a language flag as one of the columns of data uploaded.